How to Forecast Demand When Launching a New Coffee Origin

Forecasting demand for a coffee origin that customers do not yet know is difficult because historical sales do not exist. A distributor may have data for Colombian, Brazilian, Ethiopian, or Vietnamese coffee, but that does not automatically predict how Cambodian coffee will perform.

The first forecast should therefore be treated as a structured hypothesis, not as a fact.

A good launch forecast helps the distributor order enough inventory to support the market test without turning optimism into excess stock.

Forecast the product, not the country

“Demand for Cambodian coffee” is too broad to plan inventory.

Demand depends on the product:

• whole bean or ground;
• single origin or blend;
• Fine Robusta or another format;
• retail bag or hospitality supply;
• premium gift set;
• espresso product;
• online specialty release.

Start with a specific SKU, price, channel, and target customer.

The same origin can perform very differently across product formats.

Define the launch channel

A forecast for ten specialty retailers is different from a forecast for one hotel group or one national grocery chain.

List the planned channels and estimate each separately.

Possible channels include:

• specialty retail;
• e-commerce;
• cafés;
• hospitality;
• premium grocery;
• corporate gifting;
• travel retail;
• roaster wholesale.

Do not combine them into one total until each channel assumption is visible.

Use account-level assumptions first

When the market is new, account-level forecasting is more reliable than broad market-size estimates.

For each target account, estimate:

• opening order;
• expected weekly or monthly sell-through;
• likely reorder interval;
• probability the account actually launches;
• expected start date.

This forces the distributor to connect the forecast to named commercial opportunities.

Separate pipeline from confirmed demand

A retailer saying “interesting” is not the same as a purchase order.

Divide the pipeline into stages such as:

• prospect;
• sample sent;
• buyer meeting completed;
• trial approved;
• opening order confirmed;
• reordered.

Apply more confidence only as the opportunity moves closer to purchase.

This prevents a large list of prospects from being treated as inventory demand.

Build three scenarios

A simple scenario model is useful during uncertainty.

Conservative: only committed accounts launch, sell-through is modest, reorders take longer.

Base: most qualified accounts launch and sell at the expected rate.

Upside: more accounts convert, sampling works well, and reorders arrive faster.

Inventory should be planned so the business can survive the conservative case and respond to the upside without excessive delay.

Do not make the upside case the purchase order

Launch enthusiasm can cause teams to order inventory based on the most optimistic scenario.

That transfers market uncertainty into warehouse risk.

Use the base or conservative case for the first commercial order unless supply lead time makes that impossible.

If lead time is long, build a response plan for upside demand rather than simply doubling the first order.

Use comparable products carefully

The distributor may use existing sales data from similar products as a reference.

Possible comparables include:

• another emerging origin;
• a premium Southeast Asian coffee;
• a similar shelf price;
• another Fine Robusta product;
• a limited single-origin release.

The comparison is only a starting point.

Do not assume identical demand because the customer story and product experience differ.

Estimate trial conversion separately from repeat demand

An unfamiliar origin may generate strong first-purchase curiosity.

That is useful but can overstate sustainable demand.

Track:

Trial demand: first purchase driven by novelty, launch, event, sampling, or media.

Repeat demand: customers or retailers buying again because the product delivered value.

The long-term forecast should increasingly depend on repeat behavior.

Reorders are the strongest forecasting signal

A retailer opening order may reflect enthusiasm, negotiated listing, or launch support.

The second order is more informative.

Track:

• days to reorder;
• reorder quantity;
• percentage of accounts that reorder;
• whether reorder size increases or decreases;
• which SKU repeats.

After several cycles, these data can replace much of the original assumption.

Sampling can distort short-term demand

Sampling often creates a temporary sales lift.

That does not mean the new higher rate will continue indefinitely.

Separate:

• normal weekly sales;
• event sales;
• promotional sales;
• post-event baseline.

If the baseline remains higher after the event, the forecast can be revised upward with evidence.

Staff education can affect conversion

An unfamiliar origin often depends on retailer or café staff explanation.

Two otherwise similar stores may perform differently because one team understands the product and the other does not.

Record whether staff training occurred.

This helps explain performance and prevents the distributor from assuming geography or customer demographics caused every difference.

Price sensitivity should be tested, not guessed

A premium Cambodian coffee may enter a shelf price range with established alternatives.

Customers may accept the price if the product feels distinctive and credible. They may reject it if the story is unclear.

Do not immediately respond to weak sales by reducing price.

First examine:

• product fit;
• merchandising;
• staff explanation;
• sampling;
• packaging clarity;
• channel choice.

Price is one variable among several.

Forecast by SKU

If the launch includes several products, each needs its own demand estimate.

A Fine Robusta flagship may perform strongly while a gift set moves only around holidays. A filter-oriented product may do well online but poorly in hospitality.

Aggregate forecasts can hide these differences.

SKU-level planning reduces slow-moving inventory.

Include lead time in the demand model

A forecast is not only about how much will sell. It also determines when the next order must be placed.

If replenishment takes six weeks, the distributor must act earlier than if local production can replenish in several days.

Track actual lead time after every order.

The more variable the lead time, the more cautious the inventory plan should be.

Seasonality can appear even in a new category

Coffee demand may change with holidays, weather, tourism, gifting periods, retailer promotions, or hospitality occupancy.

A new origin has no direct historical seasonality, but the channels do.

Use the retailer's or category's historical seasonality as a provisional guide, then replace it with product-specific data over time.

Corporate and gift orders should be separated

One large corporate gift order can make a month look successful without creating repeat consumer demand.

Track bulk event or gift orders separately from ongoing retail sales.

They are valuable revenue, but they should not automatically raise the base forecast for normal distribution.

Do not use social engagement as a demand forecast

Views, likes, comments, or search interest can indicate awareness.

They do not equal paid orders.

Use digital signals to decide where to test demand, not how much inventory to purchase.

The forecast should ultimately connect to accounts, transactions, sell-through, and reorders.

Use a rolling forecast

Do not create one annual number and leave it unchanged.

Update the forecast on a regular rhythm, such as monthly or more frequently during launch.

Review:

• new accounts;
• lost accounts;
• actual sell-through;
• reorders;
• promotions;
• inventory;
• lead time;
• upcoming events;
• changes in product availability.

A rolling forecast learns from the market.

Track forecast error

Compare what was predicted with what actually happened.

If the forecast is consistently high, the business may be overconfident about conversion or sell-through. If consistently low, supply may be too cautious.

The purpose of measuring forecast error is improvement, not blame.

A new origin should become easier to forecast after each cycle.

Scenario example without invented market numbers

Imagine a distributor plans to launch Cambodian Fine Robusta through eight specialty retailers.

The team can model:

• number of stores likely to accept the first order;
• units per store in the opening order;
• expected weekly sell-through;
• percentage likely to reorder;
• lead time from OCC;
• one planned sampling event.

Instead of claiming a national market size, the distributor builds a forecast from the actual launch plan.

That is usually more useful for inventory decisions.

Forecast the retailer's stock as well as the distributor's stock

A distributor can mistakenly think demand is strong because retailers have purchased large opening quantities.

But if those units remain on retailer shelves, the distributor will not receive a reorder.

Where possible, collect sell-through data or at least retailer inventory feedback.

Sell-in and sell-through are different metrics.

Watch the first reorder cohort

Group accounts by launch month and track whether they reorder after a similar number of weeks.

This creates a cohort view.

If most accounts reorder within a certain range, future demand planning becomes more reliable.

If only a small share reorders, the distributor should diagnose product, channel, education, or price before expanding.

Forecasting and exclusivity should be connected

A partner asking for exclusive territory should be able to show a market development plan.

Forecasts used for exclusivity should include realistic account pipeline, purchase targets, inventory plan, and review dates.

Exclusivity should not be justified by an optimistic top-down market estimate alone.

A launch forecasting template

Track these fields for each SKU and channel:

• target account;
• opportunity stage;
• expected opening order;
• launch date;
• expected sell-through;
• expected reorder date;
• actual opening order;
• actual weekly sales where available;
• actual reorder date;
• stock on hand;
• replenishment lead time;
• confidence level.

This turns forecasting into a repeatable process.

Add an assumption register

Forecasts become difficult to improve when teams forget which assumptions produced the original number. Keep a short assumption register next to the forecast. Record the expected number of active accounts, opening order per account, base sell-through, reorder interval, promotional lift, launch date, and lead time assumption.

When actual performance differs, change the specific assumption that was wrong instead of replacing the entire forecast with a new unexplained number. This makes the model auditable and helps the distributor learn whether the main error came from account conversion, customer demand, timing, or supply.

For Cambodian coffee, an assumption register is especially useful because early market education can change conversion over time. A product that sells slowly before staff training may perform differently after tasting events and better retail explanation.

Forecast capacity as well as demand

Demand planning should be connected to the brand's ability to supply. A strong upside scenario is not useful if production, packaging, freight, or green-coffee availability cannot support it.

For each scenario, ask whether OCC and the distributor could realistically fulfill the volume while protecting freshness and product quality. If not, define an allocation rule in advance. Priority may go to existing accounts, flagship SKUs, or markets with proven reorders.

This prevents the business from treating a forecast as a promise. Forecasting should help both sides prepare for possible demand, while confirmed purchase orders and verified inventory determine what can actually be supplied.

Review forecast accuracy by channel

A blended total can hide one channel that consistently over-forecasts and another that consistently under-forecasts. Compare specialty retail, e-commerce, hospitality, gifting, and wholesale separately.

Over time, the distributor may find that Cambodian Fine Robusta has stronger repeat demand in one channel than another. That evidence should influence inventory, marketing spend, and account acquisition priorities.

Red flags

Forecast risk is high when the model is based on national population, social media attention, retailer verbal interest, or one successful event without account-level evidence.

Another warning sign is when inventory purchases increase faster than actual sell-through.

Bottom line

Demand forecasting for a new coffee origin should begin with controlled uncertainty. Define the specific product, channel, and customer. Build account-level scenarios. Separate trial from repeat demand. Track sell-through and reorders. Update the model continuously.

For Cambodian coffee, the goal is not to predict the entire market before launch. It is to create a disciplined system that learns quickly enough to scale the right products without burying the opportunity under excess inventory.

Continue to OCC Distribution for partnership planning or Contact OCC to discuss a market launch.